This study seeks to address the clinical dilemma of how much more effective, if any, a newer, more expensive form of naltrexone (Extended-Release Naltrexone, XR- NTX) is compared to the older, cheaper, oral form (O-NTX), when used in a simple, primary care-based, Medical Management protocol, and using a comparative effectiveness, cost-effectiveness, pragmatic randomized controlled trial design. If the treatment of alcohol disorders is to meaningfully expand in the era of the medical home, addiction parity, and health insurance reform, primary care physicians and payers must be convinced that effective, simple-to-use, pharmacotherapies are worth prescribing. This study will consist of a pragmatic, randomized, open-label treatment trial of 24 weeks of XR-NTX vs. O-NTX Medical Management delivered in primary care. N=234 adults >18yo with alcohol dependence will be recruited from the community to treatment at two New York City adult primary care sites in lower Manhattan, Bellevue Hospital Center and Gouverneur Diagnostic and Treatment Center. The primary outcome will be a good clinical outcome of abstinence or moderate drinking during the final 20 weeks (weeks 5-24) of a 24 week treatment trial. Cost effectiveness analysis will include long- term models of treatment impact on patient, health system, and societal costs. Secondary outcomes include other standard calendar and count measures of alcohol intake (drinks/day, days abstinent, time to first heavy drinking day), treatment retention as a continuous variable, biomarkers including GGT and CDT, and the association of mu opioid receptor Asp40 functional allele status with XR-NTX treatment effectiveness.